© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 130479
    Prev
    Next
    INDEX
    Explanations

    the verb catch

    np_acts-logits-general · gemini-2.5-flash-lite

    The main thing this neuron does is detect the word “catch” in any form or tense (catch, catches, caught, catching, etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    <strong>
    -2.20
    <em>
    -2.05
    *
    -1.90
     to
    -1.88
    t
    -1.80
     –
    -1.79
    та
    -1.78
    v
    -1.78
     和
    -1.78
    j
    -1.77
    POSITIVE LOGITS
    ꜆
    2.20
    衄
    2.16
     catches
    2.13
    ciutto
    2.13
    えもん
    2.08
     bilgiler
    1.98
    嫄
    1.98
     BOTH
    1.98
    疴
    1.97
    ˽
    1.97
    Activations Density 0.020%

    No Known Activations